Do not Resurrect That Stupid Graph
Please see my points over at File_talk:Total_bitcoins_over_time_graph.png which describe why I deleted the graph, and description of a suitable alternative. Seriously, the amount of misleading information that prior graph has disseminated into the public is horrorshow-enormous. Midnightmagic (talk) 07:11, 14 September 2014 (UTC)
Actual Supply is Different From The Tables Here
Note that the actual supply is different than what is listed on these graphs due to provably destroyed coins, duplicate coinbase, and underpays; this should be accounted-for in the graphs. Midnightmagic (talk) 16:32, 16 September 2014 (UTC)
- I'm not sure if 'destroyed' coins should be taken into account. While they may never become part of transactions again, they were at least created. Duplicate coinbase and underpays are a different matter. I'd like to take these into account in at least the description in one of my graphs (see the talk page for the deleted graph), but would also like to see some source material from which to glean actual numbers. If you, or anybody else, can help out with this, it would be appreciated. TheRealSteve (talk) 21:48, 12 April 2015 (UTC)
- I'd still like to see some examples of further technical peculiarities that make existing bitcoin unspendable, as well as examples of bitcoin that actually entered into existence when it shouldn't have (thus raising the maximum), didn't enter into existence when they should have (thus lowering the maximum) and any other case that isn't simple loss or willful destruction. There's a section on this now, so by all means, add there directly TheRealSteve (talk) 12:36, 20 July 2015 (UTC)
- The Genesis block is unspendable due to an explicit code exception. Duplicate codebases account for another.. 25 BTC or so I think, plus fees. Destroyed bitcoins like OP_RETURN slowly grow total destroyed bitcoins over time as well. You can see this difference by running: bitcoin-cli gettxoutsetinfo and subtracting the total_amount from a calculated total, from genesis. At the moment, for example, total bitcoins that can ever exist if every lost bitcoin privkey is magically found again is: 20999839.77085749 BTC.
Projected Bitcoins (Long Term) - the numbers
The Projected Bitcoins Long Term table in the page uses Bitcoin unit representation as a floating point number with an 8-digit precision to convey the actual number in satoshi, which has an integer representation inside the satoshi client.
Each of the values is based on the actual number of satoshi rewarded per block. This reward halves every 210,000 blocks. This halving presents a potential point of confusion; at some point the reward would be 610,351.5 satoshi - which at As of 4 April 2015[update] cannot exist in the satoshi client's fixed point arithmetic limit.
The value is, instead, truncated to 610,351 satoshi. The total number of satoshi generated during that reward era is thus not 210,000 * 610,351.5 = 128,173,815,000 (BTC1,281.73815000), but 210,000 * 610,351 = 128,173,710,000 (BTC1,281.73710000). The same problem presents itself with any subsequent reward eras .
An earlier version of the table in this page had a single satoshi error in some of the values, likely due to related floating point accuracy issues.
At the actual last round there should be a total of BTC20,999,999.97690000.
If using floating point math throughout, depending on the accuracy and behavior of the floating point calculations, this could end up being calculated as BTC20999999.99877760, or a difference of BTC0.02187764.
The floating point math approach will tend toward a total supply of BTC21,000,000, and a difference from the actual current maximum of BTC0.0231.
For technical purposes, the total supply should be noted as BTC20,999,999.97690 with a reminder that this depends on the current fixed point arithmetic. For colloquial use, BTC21,000,000 should be acceptable. TheRealSteve (talk) 22:01, 4 April 2015 (UTC)
Projected Bitcoins (Long Term) 'Years' column removal
I've finally removed the Years column from the "Projected Bitcoins Long Term" table. There's multiple arguments in favor of this: 1. The existing format was decimal years, which aren't very layperson friendly. 2. Laypersons may be misled to believe these year/date values are set in stone, when they are not. For relevant discussion, see the arguments regarding an earlier graph which I can only presume (as the file was deleted) was equally misleading, and this reddit comment thread (disclaimer: I'm a participant in that thread): https://www.reddit.com/r/Bitcoin/comments/31g89l/rounding_error_in_the_wiki/ . If anybody wants to put year/date numbers back in: it's a wiki, go ahead, but I'd appreciate an outline of the arguments in favor in response here :) TheRealSteve (talk) 22:11, 30 July 2015 (UTC)